Andrej Karpathy Joins Anthropic to Lead Pre-Training Research
In one of the most significant talent moves in AI this year, Andrej Karpathy — co-founder of OpenAI, former head of Tesla's Autopilot and Full Self-Driving programs, and one of the most recognized researchers in deep learning — has joined Anthropic. He started this week on the pre-training team under lead Nick Joseph.
According to Anthropic, Karpathy will build a new team focused on using Claude to accelerate pre-training research — the large-scale training runs that give Claude its core knowledge and capabilities. In his announcement on X, Karpathy said he was excited to get back to R&D while noting he remains passionate about AI education and plans to resume that work in time.
The hire comes as Anthropic's run-rate revenue has surpassed $30 billion — up from approximately $9 billion at the end of 2025 — with over 1,000 business customers each spending more than $1 million annually. Karpathy's move from the OpenAI orbit to its chief rival underscores the intensifying competition for elite AI talent.
Sources: TechCrunch, CNBC
Google I/O 2026: Gemini Spark, Omni, and the 3.5 Model Family
Google used its annual I/O developer conference to unveil a sweeping wave of AI products, headlined by three major launches: Gemini 3.5 Flash, Gemini Omni, and Gemini Spark.
Gemini 3.5 Flash
The new Gemini 3.5 Flash surpasses the previous 3.1 Pro in coding, agentic, and multimodal benchmarks — while maintaining the cost efficiency and speed the Flash series is known for, delivering 4x faster output tokens per second than competing frontier models.
Gemini Omni
Gemini Omni is Google's new multimodal generation model, capable of producing output in any modality from any input. The launch starts with video generation through Omni Flash, available now in the Gemini app, Google Flow, and YouTube Shorts. Users can generate and iteratively refine videos through conversational prompts.
Gemini Spark
Perhaps the most ambitious announcement, Gemini Spark is a personal AI agent that lives in the Gemini app and runs 24/7 on Google Cloud virtual machines — no need to keep your laptop open. Powered by Gemini 3.5 and Google's Antigravity harness, Spark can perform long-horizon tasks autonomously across Google Workspace, third-party apps, and the web. It's rolling out to trusted testers now, with beta access coming to Google AI Ultra subscribers ($100/month, down from $250) in the U.S. next week.
Sources: Google Blog, CNBC, 9to5Google
Pope Leo XIV to Present First Encyclical on AI with Anthropic Co-Founder
In an extraordinary intersection of technology and faith, the Vatican announced that Pope Leo XIV will present his first encyclical — titled Magnifica Humanitas ("Magnificent Humanity") — on May 25, alongside Christopher Olah, co-founder of Anthropic. The document centers on "the protection of the human person in the age of artificial intelligence."
The encyclical was signed on May 15, exactly 135 years after Pope Leo XIII signed Rerum Novarum, his landmark encyclical on workers' rights during the Industrial Revolution — a deliberate historical parallel. The Vatican has also launched a new AI commission to advise on the ethical implications of the technology.
The choice of Olah as co-presenter is notable: Anthropic has been in the spotlight after refusing to loosen safeguards preventing its AI models from being used for lethal autonomous warfare without human oversight or for mass surveillance. The Pentagon subsequently labeled Anthropic a "supply chain risk" — the first time such a designation was applied to a U.S. company.
Sources: National Catholic Reporter, Bloomberg, Fortune
Meta Cuts 8,000 Jobs to Fund $145 Billion AI Infrastructure Bet
Meta has begun notifying roughly 8,000 employees of layoffs — about 10% of its workforce — as part of a sweeping restructuring to redirect resources toward artificial intelligence. An additional 6,000 open positions have been cancelled, bringing the effective headcount reduction to 14,000.
The layoffs are projected to save $7-8 billion annually, partially funding Meta's staggering 2026 capital expenditure plan of $125-145 billion on data centers, GPUs, and AI development — more than double its 2025 outlay. The company has already reassigned approximately 7,000 employees to AI-focused groups.
The cuts come despite Meta reporting record quarterly revenue of $56.31 billion. More layoffs are planned for the second half of 2026 as Zuckerberg doubles down on AI infrastructure as the company's defining priority.
Sources: Tech Journal, The Next Web, Yahoo Finance
Jury Unanimously Dismisses Musk's Lawsuit Against OpenAI
A federal jury in California delivered a unanimous verdict on May 19, rejecting all of Elon Musk's claims against OpenAI and CEO Sam Altman. The nine-member advisory jury deliberated for less than two hours before finding that Musk had waited too long to file his lawsuit, with all claims barred by the statute of limitations.
Judge Yvonne Gonzalez Rogers of the U.S. District Court for the Northern District of California agreed with the jury's recommendation and dismissed the case. Musk's lawsuit had alleged breach of charitable trust and unjust enrichment as OpenAI's valuation soared, but the court never reached the merits — the case was decided on procedural timing grounds.
Musk posted on X that he intends to appeal. The decision marks a significant legal setback in his long-running campaign against the company he co-founded, though the underlying questions about OpenAI's transformation from nonprofit to for-profit remain unresolved in the public discourse.